https://gilscvblog.com/about/2022-12-23T18:15:06+00:00weekly0.6https://gilscvblog.com/2013/08/26/tutorial-on-binary-descriptors-part-1/https://gilscvblog.com/wp-content/uploads/2013/08/fith100pairs.jpgBRISK descriptor - BRISK's short pairsBRISK descriptor - BRISK's short pairshttps://gilscvblog.com/wp-content/uploads/2013/08/fourth100pairs.jpgBRISK descriptor - BRISK's short pairsBRISK descriptor - BRISK's short pairshttps://gilscvblog.com/wp-content/uploads/2013/08/third100pairs.jpgBRISK descriptor - BRISK's short pairsBRISK descriptor - BRISK's short pairshttps://gilscvblog.com/wp-content/uploads/2013/08/second100pairs.jpgBRISK descriptor - BRISK short pairsBRISK descriptor - BRISK short pairshttps://gilscvblog.com/wp-content/uploads/2013/08/first100pairs.jpgBRISK descriptor - BRISK short pairsBRISK descriptor - BRISK short pairshttps://gilscvblog.com/wp-content/uploads/2013/08/freak2.jpgfreak2https://gilscvblog.com/wp-content/uploads/2013/08/freak1.pngfreak1https://gilscvblog.com/wp-content/uploads/2013/08/brisk2.pngBRISK descriptor - BRISK sampling patternBRISK descriptor - BRISK sampling patternhttps://gilscvblog.com/wp-content/uploads/2013/08/allpairs.jpgBRISK descriptor - BRISK's short pairsBRISK descriptor - BRISK's short pairs2022-03-22T01:10:50+00:00monthlyhttps://gilscvblog.com/2018/12/18/title-to-image-search-for-improved-thumbnail-selection/https://gilscvblog.com/wp-content/uploads/2018/10/forblog2-1.pngImage search resultsGiven a query image, we compute it's embedding vector and return the images with the closest embedding vector in the embedding space. In each row, the leftmost image is the query image, the four other images are the most similar images to the query according to their embedding vectors.https://gilscvblog.com/wp-content/uploads/2018/10/forblog.pngExample of two images which are semantically similar, but extremely different in terms of the raw pixel informationExample of two images which are semantically similar, but extremely different in terms of the raw pixel informationhttps://gilscvblog.com/wp-content/uploads/2018/09/null21.pngnull2https://gilscvblog.com/wp-content/uploads/2018/09/taboola.pngTaboola's "Promoted Links" boxExamples of advertisements placed in Taboola's "Promoted Links" box. Notice that each advertisement contains both a title and an appealing image. https://gilscvblog.com/wp-content/uploads/2018/09/null14.pngImages returned by our system for the query “Don't Miss this Incredible Offer if You Fly with Delta Ends 11/8!” along with their predicted CTR.Images returned by our system for the query “Don't Miss this Incredible Offer if You Fly with Delta Ends 11/8!” along with their predicted CTR.https://gilscvblog.com/wp-content/uploads/2018/09/null13.pngImages return by our system for the query “20 Ridiculously Adorable Pet Photos That Went Viral” along with their predicted CTR.Images return by our system for the query “20 Ridiculously Adorable Pet Photos That Went Viral” along with their predicted CTR.https://gilscvblog.com/wp-content/uploads/2018/09/null12.pngImages return by our system for the query “Now You Can Turn Your Passion For Helping Others Into A Career - Become A Nurse 100% Online. Find Enrollment & Scholarships Compare Schedules Now!" along with their predicted CTR.Images return by our system for the query “Now You Can Turn Your Passion For Helping Others Into A Career - Become A Nurse 100% Online. Find Enrollment & Scholarships Compare Schedules Now!" along with their predicted CTR.https://gilscvblog.com/wp-content/uploads/2018/09/null11.pngImages returned by our system for the query “Baby Born With White Hair Stumps Doctors” along with their predicted CTR. Note that some of the rightmost images are a bit off since they are farther away from the title embedding in the joint space.Images returned by our system for the query “Baby Born With White Hair Stumps Doctors” along with their predicted CTR. Note that some of the rightmost images are a bit off since they are farther away from the title embedding in the joint space.https://gilscvblog.com/wp-content/uploads/2018/09/null10.pngImages returned by our system for the query “30 Richest Actresses in America” along with their predicted CTR. Notice this is the same title we discussed earlier when comparing our approach with the naive solution. Here we also show the predicted CTR. We assume our model mistook Steven Tyler’s for a female actress due to his long hair…Images returned by our system for the query “30 Richest Actresses in America” along with their predicted CTR. Notice this is the same title we discussed earlier when comparing our approach with the naive solution. Here we also show the predicted CTR. We assume our model mistook Steven Tyler’s for a female actress due to his long hair…https://gilscvblog.com/wp-content/uploads/2018/09/null9-e1542485225852.pngImages returned by searching in the proposed joint title-image space algorithm for the query “30 Richest Actresses in America".Images returned by searching in the proposed joint title-image space algorithm for the query “30 Richest Actresses in America". Notice the returned images are far more suitable than the one returned by the naive algorithm.2018-12-28T20:28:34+00:00monthlyhttps://gilscvblog.com/2017/01/31/emotion-recognition-in-the-wild-via-convolutional-neural-networks-and-mapped-binary-patterns/https://gilscvblog.com/wp-content/uploads/2017/01/examples.pngexamplesExamples of predictions made by our emotion classification system. https://gilscvblog.com/wp-content/uploads/2017/01/hist.pnghistHistogram of the importance of each model in the final ensemble.https://gilscvblog.com/wp-content/uploads/2017/01/results1.pngresultshttps://gilscvblog.com/wp-content/uploads/2017/01/results.pngresultsResults of our Emotion Classification method for various network's architecture and image transformations.https://gilscvblog.com/wp-content/uploads/2017/01/lbp.pnglbpLBP thresholds a small region by the center pixel's intensity to get a binary value for each pixel. https://gilscvblog.com/wp-content/uploads/2016/04/teareser_c.pngteareser_c2018-12-18T22:48:59+00:00monthlyhttps://gilscvblog.com/2015/11/07/performance-evaluation-of-binary-descriptor-introducing-the-latch-descriptor/https://gilscvblog.com/wp-content/uploads/2015/11/sfm2.png3D reconstruction results - comparing SIFT and LATCH3D reconstruction results - comparing SIFT and LATCHhttps://gilscvblog.com/wp-content/uploads/2015/11/sfm1.png3D reconstruction results - comparing SIFT and LATCH3D reconstruction results - comparing SIFT and LATCHhttps://gilscvblog.com/wp-content/uploads/2015/11/leuven2.pngRecall vs. 1-precision curves for the set "Leuven"Recall vs. 1-precision curves for the set "Leuven"https://gilscvblog.com/wp-content/uploads/2015/11/bikes2.pngRecall vs. 1-precision curves for the set "Bikes"Recall vs. 1-precision curves for the set "Bikes"https://gilscvblog.com/wp-content/uploads/2015/11/mikolajczyk-_benchmark.pngExample images from each set of the Mikolajczyk benchmarkExample images from each set of the Mikolajczyk benchmarkhttps://gilscvblog.com/wp-content/uploads/2015/11/same_not_same_examples.pngSame/not-same patchesExamples of same and not-same pair patches - the left figure presents examples of ‘same’ patch pairs and the right figure presents examples of ‘not-same’ patch pairshttps://gilscvblog.com/wp-content/uploads/2015/11/teaser_b.pngVisualization of the LATCH descriptorGiven an image patch centered around a keypoint, LATCH compares the intensity of three pixel patches in order to produce a single bit in the final binary string representing the patch. Example triplets are drawn over the patch in green and blue2018-06-16T20:37:59+00:00monthlyhttps://gilscvblog.com/2015/11/19/age-and-gender-classification-using-deep-convolutional-neural-networks/https://gilscvblog.com/wp-content/uploads/2015/11/image05.pngimage05Microsoft's how-old.net misclassification examplehttps://gilscvblog.com/wp-content/uploads/2015/11/image00.pngimage00https://gilscvblog.com/wp-content/uploads/2015/11/image02.pngimage02Microsoft's how-old.net misclassification examplehttps://gilscvblog.com/wp-content/uploads/2015/11/image07.pngimage07Microsoft's how-old.net misclassification examplehttps://gilscvblog.com/wp-content/uploads/2015/11/agemissclassificaions.pngAgeMissClassificaionsAge misclassificationshttps://gilscvblog.com/wp-content/uploads/2015/11/gendermissclassificaions.pngGenderMissClassificaionsGender misclassificationshttps://gilscvblog.com/wp-content/uploads/2015/11/cnn_illustration_a.pngcnn_illustration_aIllustration of our CNN architecturehttps://gilscvblog.com/wp-content/uploads/2015/11/teaser_a.pngteaser_aExample images from the AdienceFaces benchmark2018-06-16T20:36:19+00:00monthlyhttps://gilscvblog.com/2014/05/15/an-easy-and-practical-guide-to-3d-reconstruction/https://gilscvblog.com/wp-content/uploads/2014/05/reconstruction2.jpg3D reconstruction example3D reconstruction examplehttps://gilscvblog.com/wp-content/uploads/2014/05/reconstruction.jpg3D reconstruction example3D reconstruction example2019-06-09T10:59:23+00:00monthlyhttps://gilscvblog.com/2016/05/05/deep-learning-101-talk-at-devcon-2016/2016-05-05T09:41:17+00:00monthlyhttps://gilscvblog.com/code/2016-04-09T15:36:26+00:00weekly0.6https://gilscvblog.com/2013/08/23/bag-of-words-models-for-visual-categorization/https://gilscvblog.com/wp-content/uploads/2013/08/figure-6.jpgfigure 6https://gilscvblog.com/wp-content/uploads/2013/08/figure-5.jpgfigure 5https://gilscvblog.com/wp-content/uploads/2013/08/figure41.jpgfigure4https://gilscvblog.com/wp-content/uploads/2013/08/figure31.jpgfigure3https://gilscvblog.com/wp-content/uploads/2013/08/figure21.jpgfigure2https://gilscvblog.com/wp-content/uploads/2013/08/figure11.jpgfigure12018-03-26T12:59:58+00:00monthlyhttps://gilscvblog.com/2013/09/19/a-tutorial-on-binary-descriptors-part-2-the-brief-descriptor/https://gilscvblog.com/wp-content/uploads/2013/09/figure2.jpgfigure2https://gilscvblog.com/wp-content/uploads/2013/09/figure1.jpgfigure12022-03-22T01:11:44+00:00monthlyhttps://gilscvblog.com/2013/10/04/a-tutorial-on-binary-descriptors-part-3-the-orb-descriptor/https://gilscvblog.com/wp-content/uploads/2013/10/angle.jpgAngle calculation illustrationAngle calculation illustrationhttps://gilscvblog.com/wp-content/uploads/2013/10/figure4.jpgOrientation of the patchOrientation of the patchhttps://gilscvblog.com/wp-content/uploads/2013/10/figure3.jpgCenter of mass of the patchCenter of mass of the patchhttps://gilscvblog.com/wp-content/uploads/2013/10/figure21.jpgCenter of massCenter of masshttps://gilscvblog.com/wp-content/uploads/2013/10/figure2.jpgPatch's moments definitionPatch's moments definitionhttps://gilscvblog.com/wp-content/uploads/2013/10/figure1.jpgAn example of keypoints matching using ORBAn example of keypoints matching using ORB2019-04-13T11:37:03+00:00monthlyhttps://gilscvblog.com/2013/11/08/a-tutorial-on-binary-descriptors-part-4-the-brisk-descriptor/https://gilscvblog.com/wp-content/uploads/2013/11/comparisons.jpgBRISK descriptor - Intensity comparisonsBRISK descriptor - Intensity comparisonshttps://gilscvblog.com/wp-content/uploads/2013/11/gradient-formula.jpgBRISK descriptor - local gradients forumlaBRISK descriptor - local gradients forumlahttps://gilscvblog.com/wp-content/uploads/2013/11/figure-1.jpgBRISK descriptor - example of matching points using BRISKBRISK descriptor - example of matching points using BRISK2020-08-07T05:21:05+00:00monthlyhttps://gilscvblog.com/2013/12/09/a-tutorial-on-binary-descriptors-part-5-the-freak-descriptor/2019-06-09T10:59:13+00:00monthlyhttps://gilscvblog.com/2015/01/02/adding-rotation-invariance-to-the-brief-descriptor/https://gilscvblog.com/wp-content/uploads/2015/01/boat_1.pngRecall vs. Precision curves for the set Boat - notice that since the images depict orientation changes, the proposed rotation invariant version of BRIEF outperforms the original implementation.Recall vs. Precision curves for the set Boat - notice that since the images depict orientation changes, the proposed rotation invariant version of BRIEF outperforms the original implementation.https://gilscvblog.com/wp-content/uploads/2015/01/bikes_1.pngRecall vs. Precision curves for the set Bikes - notice that the original, not rotation invariant, version of BRIEF outperforms the rotation invariant version.Recall vs. Precision curves for the set Bikes - notice that the original, not rotation invariant, version of BRIEF outperforms the rotation invariant version.https://gilscvblog.com/wp-content/uploads/2015/01/graf_1.pngRecall vs. Precision curves for the set Graffiti - notice that since the images depict orientation changes, the proposed rotation invariant version of BRIEF outperforms the original implementation.Recall vs. Precision curves for the set Graffiti - notice that since the images depict orientation changes, the proposed rotation invariant version of BRIEF outperforms the original implementation.https://gilscvblog.com/wp-content/uploads/2015/01/image3-boat.pngBoat - zoom and rotation changesBoat - zoom and rotation changeshttps://gilscvblog.com/wp-content/uploads/2015/01/image2-bikes.pngBikes - increasing blurBikes - increasing blurhttps://gilscvblog.com/wp-content/uploads/2015/01/image1-bark.pngBark - rotation and zoom changesBark - rotation and zoom changeshttps://gilscvblog.com/wp-content/uploads/2015/01/graf.pngRecall vs. Precision curves for the set Graffiti - notice that since the images depict orientation changes, the proposed rotation invariant version of BRIEF outperforms the original implementation.Recall vs. Precision curves for the set Graffiti - notice that since the images depict orientation changes, the proposed rotation invariant version of BRIEF outperforms the original implementation.https://gilscvblog.com/wp-content/uploads/2015/01/brief_res.pngCorrect matches when using the BRIEF descriptorCorrect matches when using the BRIEF descriptorhttps://gilscvblog.com/wp-content/uploads/2015/01/boat.pngRecall vs. Precision curves for the set Boat - notice that since the images depict orientation changes, the proposed rotation invariant version of BRIEF outperforms the original implementation.Recall vs. Precision curves for the set Boat - notice that since the images depict orientation changes, the proposed rotation invariant version of BRIEF outperforms the original implementation.https://gilscvblog.com/wp-content/uploads/2015/01/bikes.pngRecall vs. Precision curves for the set Bikes - notice that the original, not rotation invariant, version of BRIEF outperforms the rotation invariant version.Recall vs. Precision curves for the set Bikes - notice that the original, not rotation invariant, version of BRIEF outperforms the rotation invariant version.2017-04-27T13:01:55+00:00monthlyhttps://gilscvblog.com/2014/06/11/136/2017-04-17T13:56:18+00:00monthlyhttps://gilscvblog.com/2013/08/18/a-short-introduction-to-descriptors/https://gilscvblog.com/wp-content/uploads/2013/08/figure7.jpgfigure7https://gilscvblog.com/wp-content/uploads/2013/08/figure6.jpgfigure6https://gilscvblog.com/wp-content/uploads/2013/08/figure5.jpgfigure5https://gilscvblog.com/wp-content/uploads/2013/08/figure4.jpgfigure4https://gilscvblog.com/wp-content/uploads/2013/08/figure3.jpgfigure3https://gilscvblog.com/wp-content/uploads/2013/08/figure2.jpgfigure2https://gilscvblog.com/wp-content/uploads/2013/08/figure1.jpgfigure12022-03-22T01:00:12+00:00monthlyhttps://gilscvblog.comdaily1.02022-12-23T18:15:06+00:00